Research of Real-time Motion Detection System in Digital Video Supervisory System

In digital video supervisory system, it's an important function that captures image sequence from video-stream and detects the moving objects in real time based on image sequence. This paper creates two processes to capture the image sequence and implement real-time motion detection. The named pipe method is used for two processes to communicate. A novel digital image processing algorithm is presented, which is fast video segmentation method based on intersection of frame differences and threshold area algorithm combined with morphological operation to extract moving region. After the moving region was obtained, the motion object can be tracked and the real-time motion detection algorithm was implemented.